JPH05507589A - No-owner frame and multiple token removal for token ring networks - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed description of the invention] No-owner frame and multiple token removal for token ring networks The sight of invention 1. Field of invention The invention as a whole uses a token-based protocol to exchange data within the ring. The field of computer networks, which consists of multiple stages of A mechanism for removing false or erroneous data or ring control information from the network. Regarding the structure.

2. Conventional technology ] According to computer networks, processors, mass storage devices, printers, etc. and various devices can communicate with each other over high-speed communication links. The above device is By being connected to, or linked to, a network, data can be passed through the station. The stations communicate with each other via communication links. connected to.

An example of such a network is a local area network (LAN). According to Ning, high-bandwidth communication between devices within a limited geographic area 0 communication links may consist of optical fibers, coaxial cables, or stranded wire bare It is normal for this to occur.

The run is configured in a ring, and the network stations are placed on one side around the ring. transmit data to. Thus one station transmits data onto the ring. If the data is sent along that ring from the station to the station run. and each station receives its data from the previous station and is relayed to the next station, and the data is ultimately delivered to the station that is supposed to receive it. In a well-known ring network, the receiver or The destination station run also relays the data to its downstream successor station, The data thus circulates around the ring to the first or source station. Ru.

These stations facilitate the orderly transfer of data over a communication link. One of the types of communication protocols that operate according to communication protocols is token recovery. It is a working system.

Such systems use tokens, special bit strings, to indicates that the application has completed its transmission.

The next Stacyrun then receives the data as per protocol instructions after receiving the token. Start sending.

In certain token ring systems, if you don't actually hold the token, the station may not be able to send data on the ring, thus If the stage ten has to transmit data along the ring, then The station removes the token from the ring data stream, i.e. ``Capture'' the token when it is received from the administrator.

The acquisition station then begins its data transfer. The stage ten is When you finish transmitting data, you transmit the token again, effectively letting it go. After that, the next station waits until one station captures it and makes another transfer. transfers the token along the ring.

Network stage ten sends and receives data in the form of frames.

Frames contain other addresses of data, error detection sequences and status flags. indicates that the frame has been received by the intended receiving or forwarding stage, e.g. The frame address, including flags, identifies the source station and the forwarding station. Separate.

All stages and attachments are identified by unique addresses. Thus , each station has associated with it its unique station address and its Contains the address of the attached device. Stations have other addresses associated with them, e.g. For example, a logical address can also be provided. Single destination from source station Frames destined for a station are thus forwarded to the destination station's address. , for each station, including, for example, the address of specific equipment attached to the destination station. the address by maintaining a list of its associated addresses. Receive a frame containing one of these.

In addition to relaying the frame to its successor station, the source stage ten also copy the system and make it available to the appropriate devices associated with that station. . The destination station also sends an applicable status flag indicating receipt of the frame. You can also set .

When the frame returns to the source stage ten, the station e.g. recognizes the source address of the frame L as its own. removes or strips the frame from the ring data stream. Take it away. After transmission, the first frame received by the destination station on the ring is Assuming the station's operation on the ring is adequate, the frames it transmits are It's lame.

The frames may change as they move along the ring. So changes affect the source address in a series of frames and transfer them to the destination There is a risk that the station will not be able to recognize it. In that case, the forwarding destination tion does not strip off the frames, these frames are continuous along the ring. The station listed as the destination station will be moved. The system will receive and process the frame repeatedly, which typically may find a copy, but it actually conflicts with or delays other processes on the station. Requires the use of a trembling stationary source.

A similar situation occurs when the stage that sends a frame onto the ring Occurs when a person becomes incapacitated before being ripped off. These frames are source addresses continues to move within the ring in the same way as the frame in which it changed. Because that If none of the active stations on the ring share that frame with their own This is because they will not recognize it as something and tear it off. Therefore, these frames, Frames that have changed responses and other frames that are continuously traversing the ring are This will be commonly referred to as a "no-owner" frame. No-owner frames are continuous There is no need to occupy ring bandwidth by continuously moving around the ring. Also, these frames must be processed every time the station receives them. If there is anything that could cause a bank log of processing at the destination station, Ru. Also, a no-owner frame is IIs and high-speed processing are often required to return to the destination station. However, the time required above is typically faster than the frame is normally received. . It is thus desirable to remove no-owner frames from the ring.

IE, No-Owner Free from an Active Ring under the EE802.5 Standard Protocol The mechanism for stripping the ring is to strip all data from the ring and then restart the ring. Reinitializing the ring, which involves resetting the ring, restarts the ring. For example, the initialization is carried out to make it operational. Claim token processing involves processing claim tokens and creating new tokens. involves an arbitration process during which the winner of the arbitration (claim token winner) ) creates a new token, reinitializing the ring takes a considerable amount of time. However, there is a rare chance that performance will be lost during this period. Thus, considering the performance, Ring reinitialization should be done as infrequently as possible.

In order to discover a no-owner frame, i.e. when does the ring need to be reinitialized? In the IEEE 802.5 standard protocol, active seven There is a stage synchronized at the beginning of each frame that it relays on the frame. Add a marker.

If that active monitor then completely traverses the ring without frame stripping, If a frame containing a marker indicating that the ring has been moved is detected, the ring is reinitialized. Start the removal process, including conversion. The above marker is in the error detection sequence of the frame. Therefore, it is not protected, thus making the removal process unnecessary for the chive in the marker. Either start the process or prevent the process from occurring.

The removal process starts from the active monitor and after placing the removal frame on the ring , strips everything from the ring it receives. Active monitor has many removal frames. send frames or send idle frames to fill the ring, while containing tokens. Keep ripping everything off the ring. Active monitor is one of the removed frames , which nearly filled the ring with removed frames and idle frames. When instructed to do so, it reinitializes the ring and issues a new token. Thus , each removal process causes the loss of all data on the ring at that time. Therefore, to minimize the loss of data on the ring, reinitialize the ring and/or There is a need for an improved removal mechanism that does not require interruption of ring processing. do.

The above removal process removes the marker from a frame, such as in the TEE802.5 protocol. Can only be used in case of ring protocol which can be placed at the beginning of . Currently, it is not possible to place a frame marker at the beginning of a frame. Protocols such as SI Fiber Distributed Data Interface (FDDI) does not develop a mechanism to handle non-owner frames, thus allowing arbitrary Can be used with frame ring protocols and/or single frame formats A removal mechanism is expected.

Each time the ring is initialized, a station is designated as the ring removal station. It is necessary to If the ring is operating according to the IEEE802.5 protocol If the same stage equation is used, i.e. the active monitor is at the beginning of each ring. After his transformation, he was named Ringbasha. If the active monitor stage ten is after When a ring fails or is otherwise removed from the ring, the ring is Must be operated without ring purge until the ring purge is completed.

Thus, an improved method for selecting a ring removal station is needed. the above The method should be executed after each ring initialization. The process that should be done between initializations should the ring varsha become inactive. It is possible.

Token rings sometimes suffer from errors such as large numbers of tokens moving around the ring at the same time. - may experience a condition. As a result of this error a'Bi, a large number of stations There is a gap that results in both sending data onto the ring at the same time. If the transmission Once the stage ten receives data frames on the ring, it removes them.

Thus, data frames sent by one stage are transmitted to the second transmission stage. - Can be removed with silane. frame before reaching the intended destination If the ring is removed, ring processing is impaired and data is lost. Official place for the ring The process may, for example, correct for multiple token conditions, and the 20th token may be removed by the button. However, if the multiple token error condition persists, , other protocol processing may also be affected and ring processing may be impaired. IE EE802.5 proto)/L/MoA11SI FDDI protocol multiplexing It has no mechanism to detect token error conditions, thus the same condition is lost. until the amount of data received impairs ring processing enough to cause ring initialization. Multiple tokens on the ring will therefore proceed unchecked. There is a need for a mechanism that can successfully detect .

ljito1nu A. A method and apparatus in the form of A. token capturing means for detecting the token and removing the token from the network; B. The token capture means sends more frames to the network after removing the token. A transmission means for transmitting the same token onto the network after sending it to the station C, When a transmission means sends one or more of the above frames to the network. D. receiving means for detecting whether the work is received on the workpiece;

by the station between the beginning of each purge process and the end of that purge process. every data frame received and the rest of the frame is stripped from the network. a purge means for performing a purge process, and the token is a token capture means. The purge means starts the purge process when one or more of the above are captured. The purge operation in progress is terminated when the receiving means detects that the above frame has been received. E, after a predetermined period of time has passed after the start of the purge process, the purge means performs each purge process. The processing is finished, and one frame is talked during the above predetermined time while the network is malfunctioning. is characterized in that it is larger than the maximum time required to propagate along the ring network. has.

Another aspect of the invention is that the station has a ring of multiple tokens and a non-owner token. 0 books on improved stages and methods that make it possible to eliminate Another aspect of light is the ability to select a station to act as a ring barsha. A method is provided.

The third aspect of the invention is an improved method for discovering and removing tokens subject to no-ownership constraints. It is the law.

One dish per plate.

A more detailed understanding of the invention may be obtained from the following description of preferred embodiments, taken together with the accompanying drawings. I hope you will read it together.

FIG. 1 is a functional block diagram of a conventional token ring network.

FIG. 2 is an illustrative diagram of a conventional information frame used in the network shown in FIG. .

FIG. 3 shows a first embodiment of the present invention that does not use sequence numbers in the parsha marker frame. As a purge station and multiple token detection/correction station according to embodiments Flowchart showing the operation of the acting station.

Figure 3A is a version of Figure 3 that includes further improvements to limit the maximum duration of the purge process. flowchart of the page method.

Figure 1 shows a preferred example of using sequence numbers in the purge marker frame for purging. A station that acts as a stage pruning and multiple token detection/correction station Flowchart showing the operation of the button.

Figure 5 shows the first example of selecting a purge station using a hello frame during purging. Flowchart showing the method.

1ms, 6A, 6B both use candidate hello frames to purge station. 6 shows a flowchart illustrating a second preferred method of selecting, in particular FIG. and the candidate stage thin procedure.

Figure 6 shows the purging station procedure, Ff! J6B is a non-candidate station Show the procedure.

Figure 7 shows how to select a purger station using candidate hello frames. 3 is a flowchart showing an improved device detection subroutine.

Figures 8.8A, 8B, 8C, and 8D are all flowcharts for an improved version of the method shown in Figures 6-6B. In particular, FIG. 8 shows a candidate stage mushroom procedure. FIG. 8A shows a purge chart. Figure 8D shows the non-candidate station procedure. The start procedure is shown.

FIG. 9 is a flowchart of a method for detecting and removing a no-owner constraint token.

Σ邂l dazzling υ1” viewing 1 smoke 1. Effects of conventional token ring Referring to FIG. 1, token ringer 0 is connected to There is one station 11, including multiple connected stations IIA-11F. or more devices 14A-14F are connected to the ring. The device 14 is a data Various types and types including processors, mass storage devices, telecom link printers, etc. can do. They all communicate information on the communication link 12 to other devices in the ring. It can be sent and received between.

Stage ten 11 transmits information in the form of frames over communication link 12. ring Communication between the above two stations may be carried out, for example, by one station 11 (e.g. I When the IA) transmits one frame to another station 11 (e.g. 11E) happen. To accomplish this, transmission stage Ten IIA forms a frame. A communication link connecting two stations with a signal consisting of a bit stream The data is transmitted to another stage IIB via the stage 12A in the direction of the arrow. transmission Station IIB, which receives the bitstream from station IIA, On the communication link 12B connecting between the stage and another stage 217110 Relay the frame.

This process continues until the frame returns to the first transmission station IIA. 0 repeated by each of the stages IIC-F.Essentially, each of the stages receiving signals via a communication link 12 connecting between the stage 9 and the preceding stage 11; , via another communication link 12 connecting it and the subsequent station 11. and send a signal. If station 11 transfers the frame to its downstream stage 5 and 11, If you are the intended recipient of the frame (e.g. 11E), A copy of the frame is maintained by one or more of the devices 14 connected thereto. so that it is processed accordingly.

Determining whether one station E1 can send a frame to the ring depends on whether the station holds the token or not. Etoken is the first transmission A special sequence of bits that stations transmit along the ring to indicate the end of It is If station 11 needs to send one frame, the talk When the token reaches the sheep station, it does not relay the token, but instead captures the token and Become an owner. Station 11 then sends one or more frames. send. After completing the frame transmission, the station run 11 follows the ring protocol. and resume the transmission of tokens along the ring. Other stations on the ring After that, it captures the token and sends the frame.

When Stage Ten 11 receives the frame it ordered in addition to the token, the frame Peel off the ring 10. Thus, one frame is transmitted along the ring. Ru. That is, the signal is relayed only once from stage 9 to the station. If the first If a station does not rip off one of its frames, the frame will move inside the ring. Therefore, the target destination stage 10 will continue to circulate the frame. Receive it many times. The time required for such a frame to travel around the ring is It is usually much faster than the average rate at which frames reach the destination stage. Ru. Thus, the destination station can use this circular frame as it normally would. Receive these frames much faster than you receive them, and process these frames more quickly. This increases the station's processing overhead and possibly This will result in an increase in station errors.

As mentioned above, information is transferred in the form of frames, whose structure is shown in Figure 2. In other words, frame 40 is divided into several fields. one frame At the beginning of the frame is a preamble field 41A and a frame containing further frame information. and a start frame delimiter field 41B indicating the beginning of the section. Ru. These fields are collectively referred to as a frame start sequence 41. Ru.

Immediately after the start of frame sequence 41, a bit frame is a token. or a frame, and if so, what type of frame. A possible frame type followed by a frame control field 42 indicating whether For example, a data frame (sending message from stage ten to stage run) (frames used for , a barge marker frame, and two fields of zero order with a blank frame. are address fields 43A-B.

i.e., a forwarding address field whose contents identify the intended recipient of the frame; DA43A and its contents are the source address field that identifies the source station. It is 5A43B. An information field 44 follows the address field 43; After itself there is a frame control field 42, an address field 43 and by the receiving station to detect errors in any of the information fields 44. Frame check and FX sequence fields including error detection sequences used Do 45 continues.

The end of a frame is indicated by an end-of-frame sequence 4G. sequence 4 6 includes an end delimiter field 46A, which is the end of one frame. and a frame status flag field 46 containing a - set of status flags. Reveal B. The status flag transfers status information to the destination station. conditioned by a station other than the source station to provide Ru. This information can be used, for example, when a frame is copied by the intended destination station. This can include whether or not an error occurred, and whether any errors were devised.

If station 11 changes the contents of frame destination address field 43, A to If it is recognized as one of its own addresses, that is, its station address or If it recognizes that it is one of the other related addresses, the attached device 14 Copy the frame without processing it. At the same time, as above (, frame Relay to the next stage run in the ring.

When a frame returns to its source station, the same stage man returns to the frame sorter. The content of the address field 43B is recognized as its own address, and the frame Strip (remove) the rest of the frame from the ring, one frame is stripped like this. When the frame start sequence 41, the frame division field 42, and and part of the address field 43 remain. These fragments form one complete frame. It is ignored by the stations on the ring because it does not work. They are Ripped off from the ring when encountering stations or purge stations.

If the source address field 43B changes as the frame moves around the ring. changes, the source station is unable to recognize the frame as its own. cannot be removed, so the source station will not rip it off. Then, the frame moves around the ring as a non-owner frame and is transferred to the destination stage. Copy processing must be performed by the station listed as the station.

Similarly, if a source station becomes inactive after transmitting a frame, The frame will not be recognized by any operating station and the frame will not be stripped off. However, these frames are also stored in the ring as non-owner frames that require copy processing. move. A ring purge mechanism is utilized to avoid such copy processing. .

2. Processing of purge marker frames without sequence numbers Ring purge of the present invention Processing is performed by selecting one or more ring purge stages to initialize the ring. executed by more stations. Figures 5 and 6 are explained below. so that at any given time only one station can act as a purge station To be elected.

While a station is functioning as a purge station, it is in “purge mode”. It is said to be in the on-1 state. Otherwise, it will be in “barge mode off” state It is in. Similarly, the purge mode for stage ten is set to 1 "on" or "on" respectively. It is called F.

The station only performs the purge process shown below in section 2 or 3 (6 purge mode on). ”Execute (in the state).

In the case of the present invention, the purge station is used during the normal processing of the ring, i.e. The purge operation is performed only if the ring exists and the ring has undergone initialization. do. The station is able to produce error-free frames, i.e. its frame sequence. such that the kens indicate that the contents on the frame did not change during transmission on the ring. Respond only to frames.

Below are 5! Purge and station selection during purge as described in ili and 6w1. Stations participating in the selection process send and receive hello frames during purge. these The frame is specified in the ANSI FDDI in the destination address field. Thus, the stage ten in the group, i.e., the All stations participating in the purging and selection process receive a hello frame during the purge. take. The specific group address used during parsha hello frames is is the predetermined group address assigned to such frames by the implementer. .

The purge station discussed below receives frames and tokens using conventional circuitry. Rip it off. Each stage Shin performs these tasks according to the ring's processing professional) S le. Execute 1 For example, if A has a station on the ring that follows the SI FDDr protocol, The system is ANSI FDDI Token Ring Media Access System (MAC) specification. ANsIX3. Time to send and receive frames and tokens according to l39-1987 During the purge process, the station performs a conventional stripping process and strip the received frame after checking the source address, or as discussed below. The entire ffI%i can be stripped from the ring.

FIG. 3 shows the flowchart of the purge process of the station according to the first embodiment of the present invention. -This is a chart. This example uses sequence numbers during the purge marker frame. Not used.

Purge stage 5 is the R (step, 5O-52) purge process after receiving the token. Start. Thus, the station receives the ring frames it receives and its Start stripping away other information.

(Step 54) At the same time, the stage cod is transmitted in the following order.

(i) the data frame it is to transmit; and (ii) one of the following as detailed below. or more of that purge marker frame, (fif) token (step 66-70) The purge station is parallel to it, as shown by arrow 53. transmission (steps 66-70), stripping (steps 54-55), and receiving. The communication (steps 56-64) process is executed. stations exist on the ring It does not distinguish between different types of tokens or priority tokens that may Therefore, if any token is received, a purge process is started.

The station receives one of its barge marker frames (step 6O-6 2) or token (steps 58 and 64) or ring initialization (steps 58 and 64). Any value greater than or equal to 0 that continues the purge process until it receives a frame indicating step 58). The verge ring is in abnormal operation, as described below. The frame is received by the purge station in the ring data stream. Indicates where it should start purging and therefore where it should stop purging do.

A barge marker frame is a barge marker frame, e.g. a frame control file. along with information identifying it as a special frame type code in field 42 (FIG. 2). purge stage ten, e.g. the stage in its source address field 43B. The barge marker frame contains information identifying the run address number (Figure 2). – An error detection sequence in the frame, preferably also containing a detection sequence. If you include an error in a frame, a station will Verge marker frames from versions are no longer interleaved with their own frames. In addition, the station may convert other frames to barge marker frames due to errors. This reduces the possibility of incorrect resolution. Redundant purge marker frame is a single Sent to avoid a purge processing crisis due to an error in the purge marker frame. It is the enemy who is defeated. Purge station transmits two barge marker frames It is desirable to do so.

Referring again to Figure 3, while the purge process continues, the purge station It strips away all that it receives. (Stage, 54-55) Purge station is a bargemer containing the address of another stage ten on the ring as discussed below. Leave a frame.

A conventional station operating normally according to the ANSI FDI) I standard protocol. The ring receives all frames received from the ring except the first frame it sent. Relay. Thus, it includes the start delimiter field 41B, frame control field field 42, destination address field 43A1 and source address of all frames. Address field 43B (FIG. 2) is relayed, and if source address field 43B (FIG. 2) is If B contains the station's address, instead of relaying it, Tear off the rest.

Similarly, the present invention performs the stripping step 54 first to perform the purge process. station receives the received frame portion up to source address field 43B. The rest of the frame can be stripped off as described above. this thing A purge station is a barge marker that contains the address of another station. It becomes possible to selectively strip frames.

The second preferred stripping step 54 allows the step tion strips all received frames from the ring in their entirety. B-chi, The station begins with a start delimiter field 41B and a status flag flag. Strip the frame ending at the end of field 46B. Thus, the purge stay purge marker frames from other stations along with purge marker frames from other stations. Removes the rest of the frame that is not removed.

During the purge process, assuming the ring process is normal, the station will first receives the frames it transmits on the network. All other frames are It must have been ripped off by the station. Thus, the purge station before the version receives one of its verge marker frames and finishes the purge process ( Strip everything from the ring (except barge marker frames that belong to another station). If taken, incorrect data and/or incorrect frames such as non-owner frames remove from the network.

During the purge process, the purge stage also receives frames for it. write and send them to ancillary equipment for stripping these frames and processing them later. send

As previously mentioned, a barge station parses when it receives any of the following: Finish processing. (i) the contained error detection sequence indicates its own bar; One of the jmarker frames is error free. (if) )-kun, or ( iii) Frame indicating ring initialization (e.g. ANSI FDDI protocol) claim token frame or beacon frame) (step 56 -62) When the barge station finishes the purge process (step 62), the regular start processing as a non-purge station controller.

Thus, a station sends or sends received frames to its downstream station. will be relayed. As part of its regular processing, the station address in the source address field 43B (FIG. 2). Instead of shipping the frame), we will instead rip it off the ring. Therefore, the purge stage The purge process ends after the application receives the first barge marker frame. And it continues to strip any redundant purge marker frames from the ring.

As explained earlier (at the beginning of the purge process, there is no purge station after that). A series of (possibly zero) barge marker frames and tokens followed by or more barge marker frames and tokens. Send a data frame. (Steps 66-70) The ring containing the data to be sent Other stations on the top then capture the tokens and transmit their data, These from other stations that eventually release (i.e. transmit) the token data frame to the barge station following the end of that barge marker frame. Therefore, it is accepted. Thus, after receiving that barge marker frame, the A station operating as a station must send these data frames. allows them to propagate completely along the ring. Thus these The data frame will contain all stations on the ring, including the destination station. After taking the bus, return to the transfer source station. Such data by purge station The appropriateness of the shipment of the marker frame is determined by the barge station receiving the barge marker frame. Reserved to terminate the purge process as soon as the purge process is completed. (Ste, Pro 0) Purge The next time the station receives a token indicating the end of the previous transmission, a new purge operation begins. start the process. (Step 5O - Under error-free processing, purge station The node is unable to receive a token because it transmitted its verge marker frame before sending the token. The barge marker frame should be received before receiving the frame.

If a station receives a token before receiving one of its barge marker frames, If taken (step), this token is passed to the step before sending the barge marker frame. must precede the token captured by the detection of multiple token error conditions. (Step 64) Multiple talk Detection activities are discussed in detail in 4B below.

If the purge station still holds one token, it , the station strips the second token from the ring and removes the second token from the ring. Continue processing. (Step 68) By stripping this second token Because the purge station can correct the current multiple token state. , the purge station does not consider this condition to be a detected multiple token error condition. The purge process ends with the receipt of the second token, and the token is replaced with the first one. Multiplexed only if received after token release and before receipt of barge marker frame Treated as a token error condition.

Similarly, if the purge station has any changes in its current purge process, If one of the verge marker frames is received before sending the verge marker frame, Strip the received barge marker frame and continue its purging process. However However, if the barge station loses its barge mark capacity during the current purge process. has already sent at least one of its verge marker frames and receives one of its verge marker frames. If the token is still protected, terminate its current purge process. . The purge station does not consider this a multiple token error condition.

3. Purge using barge marker frame with sequence number In a typical example of a purge process discussed below in Figure 4, the barge marker frame In addition to the station address and error detection sequence in the first embodiment, Purge process sequence number, including the process sequence number associated with the specific purge process. The sequence number starts with a predetermined value and the purge station starts a new purge process. is incremented each time. (Steps 52 and 65) Sent as part of the purge process Purge processing sequence where all barge marker frames contain the same sequence number When the number reaches its maximum value, restart. That is, it is wrapped in the starting value. Desire The preferred sequence number is a 1-bit number, thus alternating between °0" and 1'. purge process.

Comparing the purge process of the first embodiment shown in Fig. 3 with a typical example of purge process shown in Fig. 4. The preferred purge process then includes all the steps of the first embodiment, with two additional steps. It can be seen that adding steps 59 and 65 can be added to the pars as described above for Figure 3. The station is parallel to the transmission, stripping, and and performs reception processing.

Now, referring to Figure 4, when the barge station receives a token, a new Start the purge process. (Step 5O-52) Station Hamazuso After incrementing the purge sequence number, select “Sequence Number” for traditional reference. ゛Store the new sequence number in the register 0 Next, the station (f) the data frame it is poised to transmit, (ii)! Tanasheke one or more verge marker frames containing an instance number, (iit) talk Place the button. (Steps 66-70) Meanwhile, the stage thin from the ring at the same time while incrementing the sequence number and sending its various frames. It strips away everything it receives. (Step 54) If the station is still talking If a second token is received while holding the token, Stage 5 Abandon the object and continue its purge process without interruption. (Step 66) Thus , the station does not increment its sequence number.

A station performs its purge process by (i) one of its own purge marker frames; , (ii))-kun, or (iii) a frame instructing ring reinitialization, Continue until it receives (1'[l, purge mode remains 6 on 3) ( Steps 56-62) If the purge station is one of its barge marker frames (step 56), or an incorrect purge. Receive one or more barge marker frames containing processing sequence numbers If so (step 59), the stage tongue detects a multiple error condition. (vinegar Step 64) Multiple Token Error Condition Detection activities are discussed in detail below in Section 4.

purge stage; when a node receives one of its barge marker frames, it The purge processing sequence number in the purge marker frame is the stored sequence number. Compare with No. (Steps 69 & 59) If the sequence numbers match , which indicates that the ring is operating properly and the station Simply terminate the purge process (steps 59 & 62) if the sequence numbers If it does not match, indicating that multiple tokens are on the ring, the station Detect multiple token error conditions. (Step 59.64) 4. Multiple token error state detection activities As above, the purge station receives the token before the purge marker frame. If so, a multiple token condition is detected. Similarly, if the station is a barge marker If a sequence number is included in the frame, if the sequence number is If one or more purge Mars frames are received that contain the detect error conditions.

Once the station detects a multiple token error condition, step 64) Depending on the protocol configuration of the present invention, it does one of the following three things: (1) itself; Purge until you get the barge marker frame i! Do two things. (2) Rin immediately reinitialize the configuration. or (3) terminate the current purge process and repeat Increment the token error condition count "Purge Error Count".

If the ring protocol follows option (3), the purge station A multiple token error occurs only if it detects that state a predetermined number of times through initialization. The protocol structure of the present invention allows error conditions to be corrected if an error condition occurs within a predetermined period of time. Forcibly initializes the ring if a specified number of purge error limits are detected. or requires a certain number of consecutive detections of that state before ring initialization is performed. I might.

Using the preferred option (3), the ring is configured so that activity on the ring Correct multiple token error conditions through regular processing before being blocked by That is allowed. Similarly, the condition must be detected a set number of times. Another error condition may occur due to sporadic changes in barge marker frames. or the station does not initiate a multiple token correction procedure when causing removal. You can do whatever you like.

Rings that do not use sequence numbers in barge marker frames, i.e. the first implementation In the example purge process, the purge station has multiple token errors on the ring. The only way to check the condition is to test the condition before receiving the barge marker frame. By receiving Kun. If the station follows option (3) For example, upon receipt of a token, it increments its error condition count. if count is less than or equal to the number required for ring initialization, stage 5 is another purge process. The token crosses the ring at least twice or the token is received twice. Wait for a predetermined period of time (measured with a “purge wait” timer) long enough to , the predetermined time of the purge wait timer is 50 milliseconds, which is 25 times the required value.

This is because it is inconvenient to execute if the time is short.

The station is about to purge, and the barge marker frame on the ring crosses the ring. and return to its source station. With regular stage Shin The purge station that is running recognizes its source address in the frame and remove them. After rotation of the two tonks, the purge station The marker frame ring would have been ripped off if the station If you do not refrain from starting a barge process, mark the barge marker before each token it receives. Let's place the frame.

Finally, all tokens are associated with barge marker frames and error states would be undetectable.

If the ring uses sequence numbers, the station will If the number required for ring initialization is withdrawn, it is not enough to refrain from starting the purge process. Each time a station begins a purge process, it increments the sequence number. If there are many If the heavy token is on the ring, the station will start as soon as it receives the first token. Send a barge marker frame with a sequence number of 1 and later a second talk upon receiving the second number, it transmits a barge marker frame having the second number. first bar When the frame marker frame returns to the station, the station The error occurs because the purge number does not match the number associated with the current (second) purge process. Detect errors. Upon receiving the token, the station again issues the sequence number. The zeroth stage that increments and starts a new purge process is associated with the second token. Receive subsequent barge marker frames. Furthermore, the frame sequence number is then 77 times, and therefore the status line detects another error condition. put out If multiple tokens remain on the ring, the received barge marker The sequence number in the frame is the same as the number of the current purge process. The ring shall detect the error condition a sufficient number of times to cause ring initialization. cormorant.

Placing the sequence number in the barge marker frame ensures that the station Heavy token error conditions can be detected.

In addition, it is possible to tell the station run whether the ring has corrected its condition via canonical processing. You can make a decision. Inclusion of sequence number in barge marker frame does not require complex circuitry, i.e. the station stores the current part in memory. All you need to do is keep the number associated with the process.

If the ring uses the first mechanism, i.e. no sequence number in the marker frame. If the station uses the above-discussed If operating under option 3), the station detects a multiple token error condition. A purge wait timer is provided to determine the period of time to refrain from starting the purge process after the One of the current standard network protocols, such as AN SI FDII) r or operating stage under IEEE802-5 standard protocol. The application monitors the time it takes for the token to traverse the ring. station The token starts a timer that lasts for the standard maximum token rotation time. Thus, the The purge standby tank lasts twice the maximum rotation time to accommodate the purge process. Now you can easily maintain Purge 8! The ease of implementing the The frame circulates through the ring during periods when stations refrain from performing purge operations. need to be balanced against the possibility of Thus, the ring maker No-ownership that cycles through the ring for a limited period of time when selecting a purge method for Use a more complex purge process that includes sequence numbers in light of results that have frames. It is necessary to consider the results used.

5. Selection of purge stage tongue using parsha halo frame A purge station should be selected for each ring initialization.

Referring to FIG. 5, using the first purge station selection method, each Stations on the ring that are functioning properly are initially used as purge stations. and send one barsha halo one frame on the ring, i.e. the station address. Include the code as the source address and identify it as a Parsha hello frame. and place the containing frame. (Steps 8l-82) These stations each then contains a source address that is numerically smaller than its own address. - As a purge station according to section 2 or 3 until you receive a chassis hello frame. (Steps 85-86) The station sends a parsha hello frame. At the same time, the parsha hello frame timer is started. (Step 83 ) When the timer elapses, the station returns to another parsha halo frame. send a message. (Step 84) If at any time stage ten is purge if you receive a barge hello frame containing an address smaller than Stops acting as a station and uses its parsha hello frame timer Make it impossible. (Steps 85-86) Finally, it operates as a purge station. The station with the lowest address is the one with the lowest address. It is.

100 Mbit FDD Partial hello frame timer length for I ring It is desirable that the length is 10 seconds ± 2 seconds. The maximum length of the timer is 12 seconds.

When a station becomes a non-purge station, it contains numerically small addresses. Parsha halo to time the receipt of the next parsha halo frame. Once the frame is received, the timer is started. (Step 87) Non-barge stay Each time the controller receives such a frame, it restarts the timer. (step 88 & 87) If the parsha receives a hello frame, the parsha starts a timer. If time elapses before receiving a low frame (step 89), the non-barge The station will once again become a barge station. (Steps 80-81) Stay The application then periodically sends its parsha hello frame to Until you receive the parshah halo frame containing the dress (as mentioned in section 2 or 3) ) Execute the purge process. Thus, even if the selected purge station Even if a ring is out of service, the maximum time a ring does not have a purge station is - less than two cycles of the timer that receives the hello frame.

On the 100 Mbit FDDI ring of the timer that receives the parsha hello frame. The length of time can be 3 to 6 times the maximum length of the parsha hello frame timer. Preferably, the timer for receiving parsha hello frames has a period of 36 seconds and is The maximum time without a purge station for a purge is 72 seconds (2 timer periods). Preferably, a parsha hello frame and a timer that receives parsha hello frames. The exact period is not important, but the ratio of the periods of the two timers is. . The non-barge station is where barge stage ten sends a parsha halo frame. You need to wait 3 to 6 times the maximum time allowed. By this This prevents an inverge station from becoming a purge station unnecessarily. can. thus expediting the processing of rings with only a single purge station. can do.

Parsha hello frames should preferably include a sequence number; Before a page station sends a parsha hello frame, it sends a parsha hello frame. After incrementing the low frame sequence number, add the new number to the parsha hello frame. Incorporate it into the system.

When a non-purge station receives a parsha hello frame, the sequence number Compare the number with the number in the previous parsha hello frame. If those numbers are different, the station receives that parshahello frame as above Restart the capture timer. Conversely, if the sequence numbers are the same, the The station ignores the parsha hello frame and restarts its timer. So, if you have a ring barge station! becomes inoperable and therefore If it is no longer possible to rip off the previously sent parsha halo frame ( For example, if stage ten is removed from the ring) is the receipt of an old parsha hello frame, i.e. a frame containing the same sequence number. will not keep restarting those timers based on

timer time at which the non-barge station received its parsha hello frame When the barge station is reached (step 89), the pass from the ring Strip off any remaining purge halo frames sent by the purge station. There is a need. These frames are because the previous purge stage ten is inactive. It becomes a non-owner frame. The purge station shows this in 25. cause ring initialization (step 80) or if it is a smaller stage A non-verge station is activated by receiving a parsha hello frame containing an address. Either enable at least one purge operation to be performed before Therefore, it can be executed.

Select the station with the lowest numerical address as the barge station. Instead of choosing the highest barge stage, the ring protocol The 0 selection method that allows you to select the station that will develop the application address is Now reset only for parshah hello frames containing larger addresses They are virtually identical, except that they are Each station on the ring has a purge station. 6. The same method of selecting options must be followed. Use candidate hello frame Selection of purge station The purge station is the second option shown in Figure 6-60. The selection can be made using a suitable parsashcon selection procedure. ring After initialization (step 110), each station icon can be configured to perform the "Start" or "Restart" The start procedure (step 100) is entered. If the ring parsha is set before initialization, If the running station controller is still active, its WAS PURGER The blowfish is “true 1” indicating that it is in the final purge, and the ring is It will continue to operate as a purge station (step 111). The station controller that acquires the right to issue a token, immediately ANS Im quasi-FDD Stations that perform claim token processing according to the I-Ring protocol are Acts as Ring Purge+ when selected. (Step 112) Therefore, the ring initial There can be two parsian stations on the ring immediately after conversion.

The station acting as a ring barsha at the beginning of the selection procedure is the part shown in Figure 6A. Enter the Just 25F procedure. When you first enter a purging procedure, the station controller enters its purge mode (step 130) and purges according to Sections 2 or 3 above. Start processing.

The station sends a parsha hello frame containing a sequence number. (Step 13'2), simultaneously start the parsha halo frame, and start the next parsha halo frame. - Timing the release of the frame. (Step 132) Station is a numerical value until it receives a parsha hello frame containing a small station address. Continue processing as a ring barsha according to 2 or 3m. (Step 136 ) Only Parsha stations can send Parsha hello frames.

Other stationcons on the ring act as first candidate stations (step 114) The candidate station icon is not a single frame, but a “candidate” frame. Hello” frames are sent periodically. Referring again to Figure 6, each candidate The station controller first enters its purge mode (step 114) and either Start the purge process according to Section 3. Also, the station is the station Starts a selection timer that establishes the maximum time that a will remain a candidate. (step 114) Each candidate station is then given a station address and its candidate helloaf. Candidate Hello 2 frame containing a frame type code indicating that it is a frame emanate. (Step 16) Finally, each station has a candidate hello frame timer. , and determine when to emit the candidate hello frame. (St. Step 116) If the station console always has a candidate hello frame that contains a small address. If the program is received (step 118), the Becomes a non-candidate station. If station selection is still in progress, the station sends another candidate hello frame and re-starts the candidate hello timer. Ru.

(Step 124) If the selection timer expires and the station controller is still If the candidate is a candidate, leave the candidate station procedure and proceed to the parsha station procedure. By starting this award, it declares itself a ring purge station. ( Step 122) The Stacy Run then performs the pass ([ff16A) The second F of the embodiment will cause a hello frame to be emitted. At 100 Mbits per DDI ring, parsha hello frame timer The period of /l supplementary hello frame timer shall be 8 to 12 seconds. is desirable.

Candidate stations can be in either purging mode or non-purging mode. I can do it. When a station console becomes a candidate, it enters its purging mode. “It becomes the purging candidate 1 station.” (Step 114) 2 or 31B while the station is a candidate purge station Candidate purge step that executes processing but does not emit HelloF L, -M during purge. When a station receives a purge marker frame from another station controller, it Purge mode (i.e., stop the purge process), non-barge candidate status - Becomes Shikon. (Steps 117.119) The station then – continue to participate in selection by emitting frames. (Step 11g-1 24) Now, referring to Figure 6B, the candidate purge stage arm is located at When you receive the program, you will see a non-barge station! (step 140), stage (If it's still on), parshah The low frame sequence number register is set to minus 1, and the barsha identification register is set to -1. Set the register to a state that represents the “unknown quantity °” (Steps 142-144> After that, stage 3 starts the barsha hello frame reception timer ( Step 146), wait for receipt of another parsha hello frame, if one more time before the Virsha Hello Frame Receive Timer expires. If one parsha halo frame is not received, the stage town will receive a new parsha halo frame. The Chastity Siran selection process is started and the Parsha candidate stage 5 is reached. (St. Step 148-151> The preferred value for the hello reception timer during purge is the candidate hello timer. It is 3 to 6 times the period of . It is desirable that the hello reception timer time is 30 seconds. The new purge candidate station icon belongs to the purge station icon ahead of the ring. This means that the parsha halo frame that Ring initialization (e6B, step 1 51).

A non-purge station whose virtual identification register is set to “unknown” When a Purge Hello frame is received by a Purge Frame, its Purge Sequence Number register is and the sequence number contained in the parsha hello frame. and the source address respectively. (Steps 152-154) Then Restarts the parsha hello frame reception timer. (Step 158) The next time the station console receives a parsha hello frame, the Compare the address with the stored verger address. If the received frame If the address of the station is numerically smaller than the stored address, then the station Acupuncture cent. (Steps 160, 154) The station controller then The parsha sequence number register contained in the parsha hello frame is Reset to low sequence number and restart Virsha hello frame receive timer. Let it start. (Steps 156-158) If the non-verging station controller contains the same source address as the purging ID, If a new verger hello frame is received, a new verger hello frame, i.e. a new verger hello frame is received. If you receive a Virsha Hello frame with a (advanced) sequence number Only that Virsha hello frame reception timer is reset. (Step 1 60-164>If the 7 dresses in the Barsha Hello frame you received are purge intelligence If the address is equal to the address in another register, the station will sequence the frame. The sequence number stored in the parsha hello frame sequence number register. Compare with the can number. (Steps 162-164) If you receive Barsha Haro - If the sequence number in the frame is greater than the stored sequence number and If you indicate that the barsha halo frame you took is a “new” frame, then The station resets its sequence number register to the frame sequence number. to The station then restarts its Parsha Hello Frame Reception Timer. Start (step 158).

If the sequence number in the received Virsha hello frame is is less than or equal to the stored sequence number Stage 5 if the low frame indicates that it is not a new frame. ignores it and restarts its parsha hello receipt timer. (steps 162-164) Similarly, if the station is stored in the Virsha hello frames that contain a source address greater than the look at (Step 160) Therefore, if it acts as a ring purge station The station inside becomes inoperable and therefore links its barsha halo frame. If you do not remove the non-purge station from the old parsha halo frame, i.e., based on the receipt of frames with the same sequence number. Now there will be no words to keep restarting, if Parsha Hello receives the data When the current timeout expires, the station renews by initializing the ring. Begin the selection process.

To summarize this preferred purge stage selection procedure, candidate stations are The purge process is executed until a purge marker frame is received from the station.

Candidate stations can be continue to participate in the station selection process. Thus, they are candidates and purge We will continue to monitor the reception of Chakharov I/-M, and when non-barge candidate stations will arrive. or non-candidate (non-purge) stations. and operates as a purge station at the selected predetermined time, excluding the start time. There is only one station at most.

The selection length, i.e. the length of the selection timer, is the maximum amount of candidate hello frame timers. It is 3 to 6 times. Thus, the selection can last as long as 60 seconds.

In an alternative to this selection procedure, the candidate station does not perform a purge operation. In this alternative, steps 114.117.119 are the candidate stages shown in FIG. A 0 selection removed from the silane procedure means that the ring will remain without a purge station for a certain period of time. The process proceeds as described above except for the activity.

For example, if the station that first acts as a ring parsha is inactive (or ) The station that acquires claim token processing acts as a purge station. If immobilized, all stations act as candidates and therefore none of them Do not purge.

The selection procedure is relatively short, but during that short time the processing of the ring is performed without any active bars. At the very least, it will not be significantly affected.

The amount of time a station can be active without an active purge is limited to one selection. The maximum time is 60 seconds.

Currently, according to the purge selection procedure, the purge station cannot perform ring initialization or purge It became possible to operate within a relatively short period of time after the engine failure. Also, one station The effect that only one station acts as a purge station is ensured.

A purge mechanism with stations acting as candidates is the preferred method.

This is because at any given time there are no stations active at the ring barger station. This is because there is usually exactly one option. However, one space In order for a station to act as a candidate, it must maintain a candidate hello frame timer. Similarly, one extra selection timer is required. Barsha The selection@structure using only halo frames is simple. However, it is a given Because there is one or more stations acting as barge stations at the time of Performance decreases.

The ring maker weighs the advantages and disadvantages of each method for a particular ring and uses a barsha selection machine. These must be selected when selecting a structure.

The purge stage selection process can be performed using any of the procedures described above. at least one purge station is active on Ru. A short period of time when the varsha is inactive due to either varsha failure or ring reinitialization. There may be long intervals. This is allowed because the existence of Bhasha is is not necessary for the correct operation of the ring (for the ring in the presence of no-owner frames). For example, this short term Meanwhile, no-owner frames may result in copy processing at the destination station. The present invention attempts to minimize the time during which there is no barsha on the ring. It's something to see.

According to the mechanism of the present invention for selecting a barsha, a large number of barshas can be selected simultaneously for a short period of time. Drive on the ring. As a result of this, a large number of varshyas Performance will be slightly degraded due to the bandwidth required on the ring to transmit the system. . It is desirable to have only one barsha operational at all times.

Thus, the mechanism of the present invention operates quickly to eliminate only one of many purge stations. Only one can continue to operate.

7. Improvement of the selection algorithm during purging The following sections 8-11 are based on Section 6 and Figure 6-6B. An improved version of the purge selection algorithm shown in The presently preferred embodiments of the present invention in use include all of these modifications. 7 and 8-8D illustrate this preferred embodiment.

The above improvements provide an alternative embodiment of the invention that is intermediate in complexity to the embodiments shown in FIGS. 5 and 6-6B. Can be adapted. In particular, such an intermediate example is used in the embodiment of FIGS. 6-6B. There are new stage towns where you can omit the use of candidate hello frames. If you do not stop receiving the Virsha Hello Frame, it will prevent you from becoming a Virsha. The use of non-candidate states as shown in FIG. 6B could be maintained.

8. Ring initialization caused by failure to receive parsha hello Avoidance According to the embodiment of the invention shown in FIGS. 6 and 6-6B, "non-candidate" (i.e., If the station (in the state indicated) If a medium parsha hello frame is not received, ring initialization (step 151) is performed. ), (steps 148-150) initialize the ring. The first basis is that the lack of a barsha halo frame is a serious failure or error in the ring. This is because it is assumed that it is caused by the state. However, this assumption It was overly general.

In particular, the above algorithm results in unnecessarily too many ring initializations, each of which An individual station that would inconveniently interrupt communication during its initialization process. receives barsha halo frames that do not indicate a serious ring failure for various reasons. For example, the input buffer for that stage ten is full. , so the versi received by stage 5 while the buffer is full is Hello frames will be ignored, as another example, recently purged Stations that were active during this time may now be inactive. a new purger station can be selected without reinitializing the ring. Lol.

The method by which the improvements of the present invention dramatically reduce the number of unnecessary ring initializations is In response to a failure to receive a barsha hello frame during the hello reception 1 timer period. by making a non-candidate station a candidate station without initializing the ring. It is done.

However, if a station does not obtain a varsha selection and performs the same CLIM a certain number of times, When alternating between candidate states and non-candidate states (referred to as "rT"), link It is assumed that a serious error condition exists within the ring, and stage ten is the first Force periodization.

The implementation of this improvement is shown in FIGS. 8-8D. Step 151 of FIG. 250 and 252. In particular, the “Bajahello Receive” timer time If a non-candidate station cannot receive a barsha hello frame during (steps 148-150), a counter called CATTEMPT''' is incremented (step 250) and the station becomes a candidate station. (vinegar steps 252 and 220) Compared to Figure 6, the candidate staging procedure shown in Figure 8 (with the new new step 224.225.2 (along with step 219.220.222) 26.228.229 included.

The candidate stage has an address t- lower than the stage group's own address. When receiving a virtual hello frame or a candidate hello frame (station stations 118 and 120), the stations are implemented as previously described in Section 6 and Figure 6.6B. Stopping the purge mode (step 224) as in the example shown in FIG. In the refinement shown, the station then determines the value of the CATTEMPT force, counter. is compared with a predetermined value CLIMIT. (Step 225) The preferred value of CLIMIT is It is 5. If CATTEMF’T is not yet above the predetermined limit, the step tion becomes a non-candidate again. (Step 229) However, G ATTEM If the PT is above its limit, the station registers the CATTEMPT counter. Reset to zero (step 226), then force ring initialization. (St. Step 228) Finally, the station returns to “Start” in Figure 8D, step 228). 100), wait for ring initialization to complete before purging or becoming a candidate again . (Step 101).

If stage 5n acquires bhasha selection and becomes bhasha (step 1 18.120.122.123, Figure 8 - Point “A” in Figure 8A) Station is CA T T E M P Reset T to zero.

to (Step 237 in Figure 8A) If the improvements to the failure barsha procedure described in Section 9 below are implemented, G.A. TTEMPT is also used because the stage; must be reset to zero whenever it ceases to be purged. This is done in step 202 of FIG. 80 and step 310 of FIG.

9. Failure barsha procedure Error conditions in a ring especially involve individual stations acting as varsiers. , if the error condition is not corrected and the ring initialization is repeated, e.g. If the station has the same address as the second station, then The application involves stripping off the entire frame of the Parsha station and section will never receive those frames and therefore the end of the purge stage. will never reach , the parsha stage ten will never reach the frequency it receives. In order to strip all the frames consecutively, the ring will strip only the idle frames. It will be included.

Ultimately, this follows the traditional ANSI FDDI Token Ring protocol. TVX on that stage ten (station after transmission validity timer expires) After ring initialization, the same station The version is once again Versha, and the copy address problem is again another ring initialization. Such ring initialization would be repeated infinitely, causing Dew.

An improvement of the present invention that prevents such ring initialization iterations is shown in FIG. It is a procedure. StageSyn enters this procedure when it detects a serious error condition. (Point “C” in Figure 80) The following 10 special error conditions call the fault purging procedure. ．． This will be explained in Section 11.

When the faulty barge procedure is invoked, the station aborts its purge mode and That is, the purge process in progress is stopped. (Step 202) This makes the difference The next step is being purged.

At the same time, reset the WAS-PUARGER flag to “FALSE” and step 202), the station automatically enters purging again during ring initialization. (Step Next, the station performs the copy operations mentioned in Section 10. The station must bus the dress test (DAT). After that, the station becomes a non-candidate. (STE, BU212) As a non-candidate, Station will restart selection algorithm during purge after ring reinitialization (Step 110) or until the new barsha is replaced by a new barsha. until it stops receiving low frames and becomes a candidate station (step 14). B, 150.250.252.219.220) cannot be purged .

Since different stations become versians when the failure persier procedure is called, In particular, linkages caused by error conditions involving individual parsier stations repeated initializations are prevented.

10. Copy address test ANS 1m1EFDD I protocol is a network controller in each stage thin. Copy address performed by the controller's SMT staging management protocol Define test (DAT). For each stage 3 to run the above test DAT with one of three values: ASS, FAIL, or UNK (Unknown) FAIL contains a register that stores a parameter called means the address is the same as the address of another station on the ring . PASS means the station has been determined to have a unique address do. UNK means no judgment has been made yet.

The preferred improvement of the present invention uses DAT in two ways.

(1) A station shall Periodically check the value of DAT throughout the body.

If DAT is equal to FAIL, the station (2) Once the faulty Parsha procedure has been called, When the station passes the copy address test, i.e. Leave this procedure until you become a non-candidate until you become DAT-PASS and become a non-candidate. would not be allowed to do so.

Aspect (1) of the improvement is implemented through the fault detection subroutine of FIG. this The subroutine is executed when the station is Versha (steps 230, 236i, FIG. 8A), Candidate (steps 219 and 222 in FIG. 8) or non-candidate (step 24 in FIG. 8B) 2), it is always executed repeatedly.

Regarding Figure 7, first, the fault detection subroutine calculates the value of the DAT parameter. To check. (Step 3o2) If DAT-FAIL, the station calls the failure Virsha procedure.

(Step 304) If it is DAT-PASS or 1fDAT-UNK, The station tests whether the ring is currently active. (Step 3 06) If the ring becomes inactive (step 306), the stage The purge is stopped (step 310), and then the purge selection process is restarted. (Figure 7 step 314 of FIG. 8, step 1 l O) Aspect (2) of the improvement is a step into the failure barsha procedure described earlier in Section 9 and Figure 80. This is done by adding a drop. In particular, once the fault purge procedure is called When the station passes the copy address test, i.e. exit the procedure until the parameter value is equal to PASS (step 206). To be more precise, stage tan is not allowed until the first copy address test. After busing the DAT, the station must wait for a predetermined period of time determined by the DAT validity timer. (step 208), during which time the station should wait (step 208). Continuously checking the value of the DAT parameter (step 206.201) It is preferable that the current time is 5 minutes, if the DAT is the duration of the DAT valid timer. If it remains equal to PASS throughout, the station is in the middle of a failed purge. (step 212), but if The value of DAT reverses to FAIL or UNKNOWN at any time during this time. The entire copy address test cycle is then repeated. (Step 214) 11. Ring processing fails immediately after initialization.

9! with individual stations acting as bhashas, as described in I5. Certain error conditions within the ring can cause repeated ring initializations.

The error condition is never corrected because the same station becomes barsha after initialization. It won't happen.

In a preferred refinement of the present invention, the parsha stage is configured to handle an excessive number of rings. Discover how failures of theory live within each short time interval. This condition is detected Then, parsha stage ten calls the failed purge procedure described in Section 9 and Figure 80. This makes this station a non-candidate and a different station becomes.

In particular, if the purge station enters purging mode after ring processing has resumed, Each time the station increments a counter called P ATTEMPT. ( Step 231 of FIG. 8A) Stage Ten then converts the count value to PLIMI Compare with a predetermined number called T. (Step 233) If the count is PLI[ Step 2: When overlaying T, the parsha station calls the failed purge procedure. 35), Stop being a bhasha. Otherwise, Stacylin is purging Continue to run as.

(Step 237) While the station acts as a varsha, the varsha hello timer determines Persha hello frames are transmitted periodically at time intervals. (Step 132 ．． 136.138) The first time the barsha hello timer expires (and each subsequent time) Stacy Run resets the P ATTEMPT counter to zero. set. (Step 238) Therefore, if the station Ring initializes or becomes inactive faster than Halo 1 hour during 1 page after activation Unless interrupted by Similarly, the count cannot be calculated if the station is a candidate (step 220 in Figure 8). ), non-candidate (step 240 in FIG. 8B), or failed version (step 240 in FIG. 80). It is reset to zero each time it becomes a drop (202). Therefore, F ATTEMPT is , while the stage tongue becomes barsha after (1) the ring start amplifier and (2) when the stage tongue becomes barsha after the ring start amplifier and during the ring process is interrupted before the first barsha hello timer expires. Count the number of consecutive oscillations. The predetermined value of P LIMIT is station the maximum number of such oscillations allowed before it disqualifies itself from being purged; must be selected so that

The currently preferred value for PLIMIT is 5.

12. Purge time limit The purge process performed as described in Sections 2 or 3 is probably a purge process performed on the ring. For example, this means that the second station strips off all frames transmitted by the purge station, Occurs when the persistence station never receives its purge marker frame. In this case, the stage on the ring immediately downstream from the purge station one or more of the applications will only receive idle frames. Yes, one of these stations follows the ANSI standard FDD I protocol. and that TVX (forces ring initialization when the transmission validity timer expires) It's going to be a big deal.

The desired solution to this problem is to determine how long the station will continue to purge. The idea is to impose an upper limit on whether This is because the purge station is Start during each purge process transmitting the first barge marker frame for the process It can be executed by a timer. When the timer expires, the status even though it has not yet received its barge marker frame. , the purge process will be aborted.

Figure 3A shows the purge process described earlier in Section 2, but with a time limit for the purge process. Amended to be inclusive. The station transmits barge marker frames when it The purge limit timer starts almost at the same time as the purge limit timer starts (step 68). It's expensive. (Step 30) If the purge limit timer expires while the station If waiting to receive a marker frame (step 56.58.6) 0), the station finishes the purge process. (Ste, Pro 2.63.65) Pa The time of the limit timer is greater than the ring wait time and is the time of the TVX timer. should also be small. The preferred purge limit timer period is 2 milliseconds. Ru.

13. No-owner restriction token Certain conventional ring network protocols use a second token called a "restriction token." form a token of type . According to these protocols, the restricted token is captured only by certain privileged stages; The purpose of the restriction token is A station that has the right to communicate with other stations by sending them messages. To enable urgent messages to be exchanged without the delays caused. . When it is necessary for one privileged station to send such an urgent message, After capturing a normal (non-restricted) token, its Messenger precedes the restricted token. send sage. When the emergency communication ends, one of the privilege stages Regular ring processing is performed by removing Kun from the ring and sending an unrestricted token. resume. A restricted token is transmitted due to a privileged station failure, but is left on the ring irregularly without being replaced with a "normal" (non'H1 limited) token. There is a slight possibility. (#Limited Toe “forgotten” by its first station) Kun is a 1 non-owner.

It is called a restricted token. ) If this state continues forever, the token is captured and the There are no stations capable of transmitting messages, and the ring is limited to limited tokens and idle It will contain only the frame.

Typically, this condition is triggered by a TVX (transmission valid) timer at each station on the ring. Corrected by. According to the traditional ring protocol, the restricted token It is handled in the same way as an idle frame by the immediate process. Therefore, if one The station's TVX timer indicates that the station is in an idle frame or limited talk. If the timeout expires before any frame other than the frame is received.

The station must force a ring initialization and remove the no-owner limit token. Would.

However, if a malfunctioning privileged station that generates a no-owner restricted token If the station has transmitted one or more frames, the station failures cause these frames to be linked irregularly as "no-owner" frames. the method described in the previous paragraph will not work successfully. In particular, the existence of one no-owner one frame may cause the TVX type of the station to This prevents the master from running out of time and allows no-owner limit toggling without ring initialization occurring. Kun will be removed.

Unfortunately, the stage town during the purge as mentioned in 2-12fi also has no effect on the TVX timer. You can avoid running out of time. Especially if the parsha station is a normal talk Or! If you start the purge process every time you receive any one-time token, A parsha station receives a barge marker frame each time it receives a restriction token. will be transmitted, just like the no-owner frame mentioned in the previous paragraph, The presence of a barge marker frame on the ring causes the station's TVX timer to expire. By preventing this from occurring, ring initialization will not occur! 1iIJ limited The token would be removed.

The desired solution to both of the problems mentioned in the two preceding paragraphs is One station can perform the purge process by modifying the purge procedure The number of times it receives a restricted token without receiving a regular (non-restricted) token. count or then receive limited tokens without receiving regular tokens. Either measure the time elapsed since. Measurement of elapsed time is the limit received A predetermined limit of counting or time, which is preferable to counting the number of tokens. (preferably 10,000 counts or 1 second each), the station option will stop the purge.

The station will restart the purge as soon as it receives a regular token or a predetermined amount of time has elapsed. Open. The next predetermined time is the maximum possible time for the TVX timers of the stations on the ring. It should be greater than the period.

An alternative solution is that all stations on the ring (not just Parsha) by detecting the receipt of a restricted token that is not followed by a normal token within a predetermined time. Ru. After detecting this condition, the stage will force a ring initialization to ensure no-operation. would be to remove the limit token.

14. Recapturing the ring parsha The purge station transparently strips the non-owner frame from the ring. Regular ring processing is not interrupted; no-owner frames cause them to ring twice. before they finish traversing, i.e. from the source station to the destination station. It is ripped off after busing the source station without being ripped off.

Therefore, a frame is received by the destination station at most twice.

By using the present invention, purge stations can also handle multiple tokens on the ring. To detect. If regular ring processing corrects the condition and it persists, the station only allows it to be positively corrected by ring initialization. Hidden The purge stage ten performs purge operations without unnecessarily interrupting regular ring processing. performs processing and multiple token detection. The initial purge method performed multiple token detection. Therefore, the multiple token error condition continues unchecked for a considerable period of time, Data is lost.

The above ia discussion is limited to a series of special examples of the invention. However, the advantages of the present invention are Various changes may be made to the present invention, with particular or no exceptions being realized. It will be clear that the object of the invention is therefore within the true spirit and scope of the invention. It is intended to include such changes as may occur.

Engraving (no changes to the content) (Conventional technology) FIG, 8C FIG, 8D FIG.9 Heisei Year Month Day

Claims (1)

[Claims] 1. In a station for a token ring network, A. B. token capture means for detecting tokens received by stations on the network and removing the same from the network; B. one or more frames after the token capture means removes the token. After sending the token on the network, the token is sent on the network. C. When the above station is sent by means of transmission over the network D. receiving means for detecting when one or more frames are received; D. All data from the network between the beginning of each purge process and the end of that purge process. each time it strips the data frame and the rest of the frame received by the station. The purge means starts the purge process when one token is captured by the token capture means, and the receiving means starts the purge process when one token is captured by the token capturing means. It also terminates the barge process in progress when it detects that the above frame has been received. consisting of and E. If the purge means is discharged later than the predetermined time after the purge process starts. When each purge process is completed and the network is not out of order for the specified time above. The maximum time required for one frame to propagate along the Token Ring network at said station. 2. A type of token ring network consisting of multiple interlinking stations. A. 8. token capture means for detecting and removing tokens on the network; B. generating means for generating a set of one or more purge marker frames in response to said token capture means and transmitting said purge marker frames onto a network after said token capture means removes a token; B. A token transmission that transmits a token after the above purge marker frame is transmitted. a means of transport; D. Purge marker frame that detects the above purge marker frame on the network E. system detection means; A program that strips the rest of the frames received by the station from the network between the beginning of each purge operation and the end of each purge operation. The purge means for executing the purge process starts the purge process when one token is captured by the token capture means, and the purge marker frame detection means starts the purge process when one token is captured by the token capture means. and a station that terminates the purge process when a jimmer frame is detected. 3. 3. The station of claim 2, wherein said purge means terminates the purge process when one token is detected by said token capture means. 4. one or more data frames when they are ready for transmission 4. The station of claim 3, further comprising transmission means for transmitting a purge marker frame onto a network in response to said token capture means for removing talk, wherein said data frame is transmitted before said purge marker frame. 5. The above station further includes: A. The said set of purge marker cuffs in response to said generating means receiving one token. B. sequence numbering means for counting the number of times the purge marker frame is generated, the generating means including a sequence number in the purge marker frame; most B. sequence number storage means for storing a sequence number contained in a subsequently transmitted set of barge marker frames; D. comparison means for comparing a recently stored sequence number with a sequence number in a received purge marker; Above frame 5. The station of claim 4, further comprising: multiple token detection means for detecting a multiple token condition in response to said comparing means indicating that the program sequence number and the stored sequence number do not match. 6. A. B. counting means for counting the number of times a multiple token condition is detected; B. If the number of times a multiple token state is detected exceeds a predetermined number, ring correction processing is performed. 6. The station of claim 5, comprising means for starting the process. 7. The multiple token detection means performs ring correction in response to detection of multiple token status. 6. The station of claim 5, comprising means for starting the process. 8. 7. The station of claim 6, wherein said generating means includes information identifying a generating station in said purge marker frame. 9. 9. The station of claim 8, wherein said purging means refrains from stripping purge marker frames containing information identifying another station from the network. 10. The station according to claim 9, wherein the comparing means compares the sequence numbers in the purge marker frame only if the frame includes information identifying a station. tion. 11. In a computer network station, A. B. token capturing means for detecting and removing tokens on the network; B. B. transmitting means for transmitting one or more data frames onto a network in response to token removal of said token capturing means when said data frames are ready for transmission; generating means responsive to said transmission means and said token capture means, wherein said token capture means removes one token to generate a set of one or more purge marker frames containing an address identifying a station; The generation means generates a sequence number equal to the number of times the set has occurred, and the generation means transmits the original purge marker frame onto the network either after the token capture means removes the token or after the transmission of the data frame is completed. D. Talk after the above set of purge marker frames are transmitted E. token transmission means for transmitting a token; F. sequence number storage means for storing sequence numbers included in the last set of purge marker frames released; Net purge marker frame G. detection means for detecting on the workpiece; H. comparison means for comparing the sequence number in the detection purge marker frame containing the station's address with the stored sequence number; Talk in response to the token capture means and purge marker frame detection means. A token or purge marker frame containing an address that identifies another station before the signal is captured and an error-free purge marker frame or token is detected. I. purge means for stripping all frames and the remainder of the received frames from the network, except for frames; I. 1. The above frame sequence number and each stored sequence number match. 2. the above comparison means showing that there is no such thing; One of its purge marker frames is detected the above token detection means for removing the token before it is detected; multiple token detection means for detecting the state; K. counting means for counting the number of times a multiple token condition is detected; and means for performing a ring correction process when the number of times a multiple token condition is detected exceeds a predetermined value. 12. How to remove exogenous or false information from computer networks A. Detecting a token on the network;B. remove the same token; c. A set of one or more purge markers in response to the removal of a token. D. generate a purge marker frame and transmit the purge marker frame onto the network; Release the above tokens to the network and remove all data frames or received data until an error-free purge marker frame or token is detected from the E-network. Stripping off the remainder of the taken frame. 13. The above steps further include A. Counts the number of times the above purge marker frame set is transmitted on the network. Toshi, B. The number is included in the purge marker frame, and C. transmitted purgemer Store the number contained in the last set of frames;D. Compare the store number with the number in the received purge marker frame; E. If the above numbers do not match, detect a multiple token condition, and F. G. Count the number of times a multiple token condition is detected; 13. The purge method according to claim 12, comprising the step of starting ring correction processing when the number of times a multiple token condition is detected exceeds a predetermined value. 14. The generation step of the purge marker frame above is included in the frame information. 14. The method of claim 13, comprising identifying the 15. The above stripping step includes a purge mark containing information identifying another station. 15. The purging method of claim 14, which refrains from stripping off the car frame. 16. 16. The purge method according to claim 15, wherein the step of comparing sequence numbers compares sequence numbers only when the purge marker includes information identifying a station. 17. 17. The method of claim 16, wherein the method further includes the step of transmitting the data frame over the network after the token is removed and before the set of purge marker frames is transmitted. 18. In a computer network with multiple mutual ring stations A. in a method for removing exogenous and erroneous information of the type including data frames, no-owner frames and multiple tokens from a computer network. Net Capture and detect one token received by one station on the network, and B. After the token is removed, the token is transmitted over the network after transmitting further frames onto the network, and C. Stacie one or more of the above files sent by means of transmission over a network. receiving and detecting when a frame is received;D. The purge means is received by the station between the start and end of each purge process. The purge process is performed by using a purge method to strip all data frames and the remainder of the frame from the network each time the purge process starts when a token is captured, and the purge process starts when a token is captured, and the purge process Terminates the purge process in progress when E. said method, wherein each purge process ends no later than a predetermined time after the start of the purge process, and said predetermined time is greater than a large amount of time required for one frame to propagate through the entire token ring network when the network is not faulty. . 19. A. B. detection means for detecting network initialization; The purger hello frame periodically includes a purger hello frame containing an address that identifies the generating station, and transmits the purger hello frame onto the network in response to network initialization. - frame generating means; C. D. purger hello frame detection means for detecting a purger hello frame; The address and station address contained in the detected purger hello frame A comparison means for comparing dresses; and E. If the address in the detected purger hello frame is numerically greater than the station address, the purger hello F. means for setting a timer for receiving frames; Before the timer expires to receive the above purger hello frame, a station address that is numerically smaller than the station address a token ring station that includes a means for enabling the station to start the purge process if the station does not detect a purge hello frame containing the station address; tion. 20. The station generating means causes the station to generate the purger harrow frame. address that is numerically smaller than its own address before the program receive timer expires. If a purger hello frame is received that contains a 20. The station of claim 19, comprising means for refraining from generating a system. 21. In a computer network consisting of multiple mutually linking stations. A. detecting network initialization; B. Purjah loaf containing the address corresponding to the station upon initialization C. frame is periodically generated and the purger hello frame is transmitted onto the network. Detects purger hello frame and D. Compare the address in the detected purger hello frame with the address corresponding to the station, and E. If you receive If the taken frame address is numerically greater than the station address, interrupt the generation of the purger hello frame; F. Set the purger hello frame reception timer, G. If station a A parjar hello frame containing an address numerically larger than the address is tied above. If it is received before the timeout of H. Occurrence status I. enabling a station to perform a purge operation if a purger hello frame containing an address numerically greater than the specified address is not received before expiration of said timer; Said method comprising the steps of repeating steps A-H. 22. In computer networks with multiple interlinking stations How to operate the station when selecting a network purge station By law, A. detecting network initialization; B. Set the purger selection timer; C. D. periodically generating a candidate hello frame including the address of the generating station; transmitting the candidate hello frame onto the network; D. Detects candidate hello frames and scans the addresses contained in the detected frames. E. The address in the detection frame is the station address. If it is numerically smaller than the dress, the station generation means is used to generate a candidate hello frame. In order to prevent the shipment of F. If the candidate halo has a smaller address - declares the station as a network purge station if the purge selection timer expires before the frame is received by the station. Ru. Said method comprising the steps. 23. The above method further comprises: A. Detect purger hello frame, B. Purjahello frame receiving tie Set C. D. Set the purger hello sequence number to a predetermined value. pa - Set the Jahalo identification value to a predetermined value, and F. Compare the identification information in the detected purger hello frame with a predetermined purger hello identification value; If the frame value is larger than the predetermined value, the purger harrow separation value is detected at the frame value. Set the purger harrow identification value in the G. If the frame identification value and If the station perger hello identification value does not match, the sequence number in the detection frame is compared with the station perger hello sequence number, and the H. If the frame identification value is greater or the frame sequence number and station purge If the parja hello sequence number matches, the parja hello frame reception Reset now, I. If the station purger has a frame identification value greater than the hello identification value or a frame with a matching identification value and higher sequence number, then A raw frame is received by the station before the timer on which it is received expires. 23. A method of operating a purge station as claimed in claim 22, comprising the step of restarting purge station selection if the purge station selection is not taken.